Operational issues:

• lbprod DB outage at CERN yesterday afternoon
  • Caused massive job failures across the grid
• High failure rate for MCFastSimulation jobs
  • Jobs incorrectly asses how much time they need to execute
  • Fix applied, but it only affects new productions
  • Some old ones are still present in the system, so we have to wait for them to disappear
    • Failure rate is still a bit high, but gradually decreasing